# Outreach pack: search-docs-for-ai-agents > **This is a template-quality pack.** It gives you the structure, voice rules, message templates, and objection handling for the first 50 outbound conversations. To get hand-personalized prospects with real LinkedIn URLs and detail-specific opening lines (the AfterHours-style format), commission a Phase A engagement: [/factory/operator-partnership/](/factory/operator-partnership/?from=search-docs-for-ai-agents) ICP: Precise. Terse. Composable. Copy reads like good API documentation: no filler, no enthusiasm performance, no jargon for its own sake. Sentences are short. Nouns do the work. If a paragraph can be a table, it is a table. Live URL: https://wishdeal.com/factory/builds/search-docs-for-ai-agents/ GTM motion: Developer-led free tier → usage-based upgrade, distributed via AI Discord communities, GitHub integrations, and Hacker News Show HN launch targeting LangChain/CrewAI/AutoGen builders. Voice rules: precise, terse, composable. No em dashes. No AI jargon. Lead with a specific detail you can verify in 90 seconds of looking at their company. --- ## Where to find your first 50 prospects Pick ONE distribution path for the first batch. The mistake is running three half-funded channels. 1. **LinkedIn Sales Navigator search** for the ICP title at companies in the size band. Save the search. Pull 100 contacts, top 50 get the personalized first-touch. 2. **Two industry-specific Reddit / Slack / Discord communities** where your ICP hangs out. Lurk for a week. Comment for two weeks. Pitch in week three. 3. **A first-party content asset** (one-page tool, one ROI calculator, one open-source script) that the ICP would link to. Promote in two communities and one industry newsletter. Pick one. Run for 30 days. Measure response rate at touch 1, touch 2, touch 3 separately. --- ## LinkedIn first message (template - 280 to 295 chars) > Hi {first_name}, {specific_detail_about_them_or_their_company}. The thing about {their_workflow} is {the_specific_failure_mode}. {Open_question_about_how_they_handle_it}? Curious to compare notes. **Replace before sending:** - `{first_name}` - their actual first name (no "Mr/Ms") - `{specific_detail_about_them_or_their_company}` - something you found in 90 seconds (a recent post, a hire, a press release, a review) - `{their_workflow}` - the exact internal workflow this product addresses - `{the_specific_failure_mode}` - the named pain (from the dossier) - `{Open_question_about_how_they_handle_it}` - never closed-ended, always invites a comparison --- ## Email follow-up (template - 130 to 150 words) > Subject: {specific_pain_phrase} > > Hi {first_name}, > > Quick follow-up. {Their_company_or_workflow_specific_observation}. > > {Product_name} {one_sentence_what_it_does}. {One_sentence_what_changes_for_them_specifically}. > > Worth 15 minutes for a live demo? I will use {specific_scenario_from_their_world}. > > https://wishdeal.com/factory/builds/search-docs-for-ai-agents/ > > Wes **Replace before sending:** - `{specific_pain_phrase}` - 4 to 6 word phrase that sounds like internal slack, not marketing - Keep it under 150 words. Long emails feel like newsletters. --- ## Second touch (template - 90 to 100 chars) > {first_name}, quick bump. Still on the {specific_pain_phrase} question at {company}? Wes --- ## Third touch / breakup (template - 60 to 80 words) > Subject: should I stop? > > Hi {first_name}, > > Two messages, no reply. I am not going to keep poking. If this is not the right time, no problem. If it is the right problem but I picked the wrong angle, reply with one word and I will try a different door. > > Either way, the resource is here whenever you want it: https://wishdeal.com/factory/builds/search-docs-for-ai-agents/ > > Wes --- ## Cold-call script (45-second open) If you are calling, the first 15 seconds is everything. > "Hi {first_name}, this is Wes. I will be brief - I work on a tool for {ICP} that handles {specific_workflow}. The reason I am calling you specifically is {specific_detail}. Before I waste your time, can I ask one question about how you handle {workflow} today?" If they say yes, ask the question and shut up. If they say no or push back, this: > "Totally fair. Last thing: is this a problem you would even consider buying for, or is it baked into how you operate today? If it is baked in, I will not bother you again." This works because it gives them a clean exit, which paradoxically creates a real conversation about a quarter of the time. --- ## Likely objections (and handled responses) These are the patterns most likely to come up based on the skeptic memo for this idea. Have a one-sentence answer for each before you start outbound. - **Tavily already owns developer mindshare for AI-agent search and has raised $6M — zero switching cost means any price adv** - **OpenAI and Anthropic are embedding native web search directly into their APIs, which could eliminate the independent sea** - **Underlying search infrastructure costs (Bing API, Common Crawl licensing, or custom crawl infra) spike unpredictably and** - **Developer tools have near-zero lock-in — a single competitor GitHub README mentioning lower cost per query can trigger c** --- ## What good looks like at 50 sends - 25 to 35 percent response rate at touch 1 (any reply, including "no") - 5 to 10 booked discovery calls - 1 to 3 paid pilots starting in week 4 - A list of objection patterns you did not expect, that update the dossier If you are below 15 percent reply rate at touch 1, the ICP is wrong or the personalization is not specific enough. Pause. Rewrite the first-message template. Send 10 more before scaling. --- ## What this pack does NOT include - Hand-personalized prospect rows with real LinkedIn URLs (that is the Phase A operator-partnership deliverable) - Verified phone numbers or email addresses (use Apollo, Clay, or an enrichment vendor; we don't ship contact data) - Tracking pixels or auto-send infrastructure (this is a copy-paste pack, never auto-send) - Compliance review for regulated verticals (legal/medical/financial - get your own review) --- *Template generated from brand brief, financial model, and skeptic memo. To get a hand-personalized 15-prospect version like AfterHours, commission a [Phase A engagement](/factory/operator-partnership/?from=search-docs-for-ai-agents).*